This paper devotes to the image compression and encryption problems. We develop a novel hybrid scheme based on block compressive sensing. Concentrate on taking full advantage of the different frequency coefficients sparsity, the nonuniform sampling strategy is adopted to improve the compression efficiency. First, the discrete cosine transform coefficients matrices of blocks are transformed into vectors by zigzag scanning. The different frequency components are extracted in the front, middle, and back of vectors, respectively. Using the measurement matrices with different dimensions, the combination of lowand high-frequency components, together with the medium-frequency coefficients are compressed simultaneously. Second, the recombinational block measurements are re-encrypted by the permutation-diffusion framework. The logistic map is introduced for key stream generation. In order to accomplish a sensitive and effective cryptosystem, the control strategy for secret keys is employed. The simulation results indicate that the proposed scheme forms a high balance between reconstruction performance, storage and computational complexity, and hardware implementation. Moreover, the security analyses demonstrate the satisfactory performance and effectiveness of the proposed cryptosystem. The scheme can work efficiently in the parallel computing environment, especially for the images with medium and large size. INDEX TERMS Block compressive sensing, image cryptosystem, logistic map, nonuniform sampling strategy.
With the advent of the physical memristor, various memristive neural network models have been designed and analyzed to mimic some human brain functions. However, there is a realistic issue because many works reported the coupling of neuron models using either memristive synapse or memristive autapse, whereas in the real brain, a neuron can interact with both another neuron (memristive synapse) and with itself (memristive autapse). Two main ideas are developed in this work. First, we investigate the dynamics of two different neurons coupled via memristive synapse and memristive autapse. The analyses indicate that the global dynamics of this highly relies on the neuron's coupling strength. Second, a cryptographic scheme based on both S-box driven block compressive sensing and the memristive autapse synapse model is proposed. Performance analyses indicate that the coupling strength of the proposed neural network model can be adjusted to increase or decrease the security of medical data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.